Lattica, an FHE-based platform enabling safe and personal use of AI within the cloud, has emerged from stealth. The corporate is backed by Konstantin Lomashuk’s Cyber Fund, and Sandeep Nailwal, co-founder of Polygon Community and Sentient: The Open AGI Basis, amongst others.
Lattica’s know-how represents a important new customary for industries resembling healthcare, finance and authorities sectors, the place knowledge privateness and safety issues have restricted AI adoption. In keeping with Cisco’s 2025 AI Briefing: CEO Version, 70% of CEOs surveyed admitted caring in regards to the state of their networks because of the rising adoption of AI, with 34% citing safety as a serious barrier to adoption.
Totally Homomorphic Encryption (FHE) – the “holy grail” of cryptography within the final decade – ensures all communication between AI suppliers and finish customers stays encrypted, with no need to decrypt it, however on account of longstanding computational inefficiencies, FHE has but to be broadly adopted. By capitalizing on the most recent breakthroughs within the AI acceleration stack, Lattica leverages superior acceleration methods to operationalize FHE.
Led by founder and CEO, Dr. Rotem Tsabary, who holds a PhD in lattice-based cryptography from the Weizmann Institute of Science, Lattica takes benefit of the foundational mathematical similarities between FHE and machine studying to supply a hardware-agnostic, cloud-based platform that makes use of FHE to ship safe and personal use of AI.
A key differentiator powering Lattica’s resolution is its Homomorphic Encryption Abstraction Layer (HEAL), which boosts FHE efficiency and standardizes its acceleration. A cloud-based service, HEAL serves as a common bridge connecting FHE functions and AI algorithms throughout a various vary of {hardware} together with GPUs, TPUs and CPUs, in addition to devoted accelerators like ASICs and FPGAs.
“By combining the developments of {hardware} acceleration with software-based optimization, we realized that not solely might we enhance FHE effectivity to the purpose of business viability, however use it to unravel important knowledge dilemmas holding again AI’s adoption in delicate industries,“ stated Dr. Rotem Tsabary, founder and CEO of Lattica. “We’re enabling sensible FHE by creating an answer that’s tailored for neural networks.”
As a part of its emergence from stealth, Lattica has made demos of the platform out there on its web site, alongside insights from an in-depth survey inside the FHE group. Survey outcomes validate Lattica’s method, revealing {that a} majority (71%) of respondents consider FHE adoption will likely be achieved by a mix of {hardware} and software program.
“Lattica is pushing the boundaries of Totally Homomorphic Encryption, fixing one of the important challenges in AI safety,” stated Konstantin Lomashuk, Managing Associate at Cyber Fund. “Cyber Fund is proud to have led Lattica’s pre-seed spherical. That is the form of deep-tech innovation that defines the long run, and we’re excited to see Lattica main the best way.”
Lattica’s concentrate on healthcare and finance additional underscores the platform’s relevance, with potential functions in safe knowledge evaluation for medical analysis and encrypted monetary transactions.
“Lattica’s product-first method essentially transforms delicate knowledge processing within the AI ecosystem,” stated Sandeep Nailwal, co-founder of Polygon Community and investor in Lattica. “Lattica has made FHE a actuality that’s each sensible and scalable, as Tsabary and her analysis staff is proving that advances within the machine studying stack can considerably enhance the efficiency of FHE and have a right away influence available on the market.”
About Lattica
Lattica permits querying AI fashions with Totally Homomorphic Encryption, providing FHE as a hardware-agnostic, cloud-based service. The platform leads in scientific innovation by making certain that consumer queries stay encrypted all through all the machine-learning inference course of.
Lattica’s Homomorphic Encryption Abstraction Layer (HEAL) connects FHE functions, algorithm implementations, and various {hardware} backends, making safe AI computation as accessible as conventional cloud-based AI providers. The corporate is headquartered in Tel Aviv.
